whats the role of maxrate and bufsize in h264 codec

[Anonymous543]

i wanna learn how to do Constrained encoding (VBV / maximum bit rate) encoding
but i dont understand maxrate and bufsize
i saw this post but its still confusing
https://superuser.com/questions/945413/how-to-consider-bitrate-maxrate-and-bufsize-of-...-video-for-web

[blud7]

From the ffmpeg wiki:

-b:v: specifies the target (average) bit rate for the encoder to use
-minrate specifies a minimum tolerance to be used
-maxrate specifies a maximum tolerance. this is only used in conjunction with bufsize
-bufsize specifies the decoder buffer size, which determines the variability of the output bitrate

Think of maxrate as an upper limit on bitrate as in you don't want to go above it.
bufsize specifies how often you compare the current bitrate to the desired average bitrate (-b:v).

Consider this command line from the wiki as well:

ffmpeg -i input -c:v libx264 -b:v 2M -maxrate 2M -bufsize 1M output.mp4

-b:v 2M specifies the average bitrate
-maxrate 2M says don't exceed 2M (avg) when encoding
-bufsize 1M tells the encoder to check every 1M to see if the current bitrate matches the average bitrate

[Anonymous543]

Shall I use this one Constrained encoding (VBV / maximum bit rate) rate control mode or 2 pass rate control mode or any other mode(suggest) for rencoding movies and other multimedia

Am..thinking to make a website for movies and other multimedia contents ,where people can download good quality movies and other multimedia in good quality at low file size..

So which rate control mode will be good?
Am targeting filesize(bitrate) that's why am not considering crf or constant qp modes

[jagabo]

I don't see any reason to use VBV encoding for downloadable files. Or bitrate based encoding. Why should a movie of mostly still shots and little noise (ie, easily compressible) be encoded with the same bitrate as another movie with lots of action, grain, smoke, flickering lights, etc. (ie, difficult to compress)? If you use high bitrates for all you're wasting bitrate on the easily compressed movies. If you use low bitrates for all the action movies are going to look bad. But if you must use bitrate based encoding you should use 2-pass encoding. Single pass bitrate based encoding delivers lower quality overall.

[Anonymous543]

I don't see any reason to use VBV encoding for downloadable files. Or bitrate based encoding.

for reasonable download file size and quality

Why should a movie of mostly still shots and little noise (ie, easily compressible) be encoded with the same bitrate as another movie with lots of action, grain, smoke, flickering lights, etc. (ie, difficult to compress)? If you use high bitrates for all you're wasting bitrate on the easily compressed movies.

i will change bitrates according to action and video resolution

But if you must use bitrate based encoding you should use 2-pass encoding. Single pass bitrate based encoding delivers lower quality overall.

with -b:v we specify only target filesize i guess..so may be what if it increases too much during complex sections? what if bitrates goes much higher(during playback) then hardware's read write speed or processing speed?

btw am asking this questions because i am lil curious.. in current genration may be this kind of situations dont happen much or rarely happens

[jagabo]

I don't see any reason to use VBV encoding for downloadable files. Or bitrate based encoding.

for reasonable download file size and quality

You don't need bitrate based encoding for that. CRF and 2-pass VBR deliver about the same quality when the bitrates match.

Why should a movie of mostly still shots and little noise (ie, easily compressible) be encoded with the same bitrate as another movie with lots of action, grain, smoke, flickering lights, etc. (ie, difficult to compress)? If you use high bitrates for all you're wasting bitrate on the easily compressed movies.

i will change bitrates according to action and video resolution

That's exactly what CRF encoding is for. You specify the quality you want and you get exactly the right bitrate needed to deliver that quality.

But if you must use bitrate based encoding you should use 2-pass encoding. Single pass bitrate based encoding delivers lower quality overall.

with -b:v we specify only target filesize i guess..so may be what if it increases too much during complex sections? what if bitrates goes much higher(during playback) then hardware's read write speed or processing speed?

btw am asking this questions because i am lil curious.. in current genration may be this kind of situations dont happen much or rarely happens

VBV is really only an issue with disc based playback -- DVD and Blu-ray. The limited transfer rate from the disc to the decoder, and the limited amount of memory within the player will cause problems if the bitrate spikes too high for too long. But that's usually not an issue until you get into Blu-ray bitrates (20 to 30 Mb/s). If you're shooting for small file sizes that's not going to be an issue for you.

[Anonymous543]

VBV is really only an issue with disc based playback -- DVD and Blu-ray. The limited transfer rate from the disc to the decoder, and the limited amount of memory within the player will cause problems if the bitrate spikes too high for too long. But that's usually not an issue until you get into Blu-ray bitrates (20 to 30 Mb/s). If you're shooting for small file sizes that's not going to be an issue for you.

Thanks for explaination

[pandy]

VBV and max bitrate are important also for broadcast.